Differential scanning calorimetry studies

Germany). The clear supernatant was collected and DOX was quantified by RP-HPLC using an Agilent 1100 Series unit (Agilent Technologies Deutschland GmbH, Böblingen, Germany) equipped with an UV-detector and a fluorescence detector model Spectra System FL 3000 (TSP Thermo Separation Products GmbH, Egelsbach, Germany). DOX was detected by its fluorescence at an excitation wavelength of 480 nm and an emission wavelength of 550 nm and by its UV-absorbance at 254 nm. The amount of DOX was quantified by the area under its characteristic fluorescence peak at 8.1 min retention time. Mobile phase and measurement setup were identical with the described under “3.3.2 Measuring the DOX concentration”, p.84.

4. Results and Discussion

4.1. Pre-formulation studies on doxorubicin-loaded anionic

DPPC DPPE DPPG DOX Peak area,

[kJ/mol]

Peak height, [kW/mol]

Peak area, [kJ/mol]

Peak height, [kW/mol]

Peak area, [kJ/mol]

Peak height, [kW/mol]

(–) 31.451 ± 0.050 1.181 ± 0.013 30.472 ± 0.130 1.026 ± 0.006 41.001 ± 0.1162 1.174 ± 0.018 (+) 30.538 ± 0.092 0.879 ± 0.007 29.456 ± 0.063 0.715 ± 0.004 34.464 ± 0.042 0.529 ± 0.021 Table 9: Differential scanning calorimetry data on the phase transition behavior of various phospholipids alone and upon interaction with doxorubicin (n = 9). Numerical data established is closely compliant with previous work [179].

In the case of DPPE, the main transition peak was centered at 64.31 °C for the pure phospholipid and was slightly moved by less than 1 °C to 63.84 °C upon addition of DOX.

The addition of DOX to DPPE caused similar moderate effects as in the case of DPPC – a slight peak area and peak height decrease and peak broadening.

The same thermotropic phenomena but much more pronounced were observed in the case of DPPG, too. For the pure DPPG the Tm was found at 39.84 °C and sank by 2.16 °C after addition of DOX, while the Tm depletions for DPPC and DPPE were only correspondingly 1.01 °C and 0.47 °C.

The reduction of peak height caused by DOX was 54.94% for DPPG vs. 30.31% for DPPE and 25.57% for DPPC. Regarding the peak area decrease upon addition of DOX, nearly the same progression was demonstrated by the three phospholipid candidates, too: 15.94% for DPPG, 3.33% for DPPE, and 2.90% for DPPC. Phospholipid-free DOX solutions, used as negative controls, showed no transition in the temperature range from 20 ºC to 85 ºC at the given concentration.

Based on the thermotropic modifications of the selected phospholipid candidates caused by their more or less intensive interaction with DOX, DPPG was selected as the phospholipid with the highest complexation affinity amongst them. These findings are also consistent with the calorimetric studies of Constantinides et al. (1986) [179], who found none or a very minor transition changes in the case of DPPC, but a remarkable effect of DOX over the behavior of DPPG. In the present study the phospholipid transition modifications in DOX-DPPC and DOX-DPPE systems were significant, apparently owing to the up to 17-fold higher DOX / phospholipid ratios and the 10-fold lower electrolyte concentrations used.

Nevertheless, the substantial difference in the change of the thermotropic behavior of DPPG after complexation with DOX was still apparent compared to the slight alterations, demonstrated by the zwitterionic phospholipids DPPC and DPPE.

Numerous authors have also demonstrated the prevailing role of strong electrostatic interactions and the auxiliary contribution of hydrophobic forces in the formation of complexes between DOX and anionic phospholipids [166, 170, 181]. Since DOX is one of the most hydrophilic anthracyclines (octanol / water partition coefficient of 1.1 [182]) its complexation affinity to zwitterionic phospholipid layers e.g. DPPC is expectedly low as demonstrated above.

Interestingly, the affinity of DPPE to DOX was constantly higher than DPPC in spite of the large similarity of their chemical structures. This can be attributed either to the different lyotropic equilibrium phases (Lα for DPPC and HII for DPPE) or to the ability of DPPE to form hydrogen bonds with DOX and therefore - a more stable complex than DPPC.

Complementary CLSM data have shown that mixed DPPC-DPPE bilayers can indeed complex DOX to some extend (data not shown) which possibly occurs through the interplay of hydrogen bonding and hydrophobic interactions.

In the second experiment series the phase behavior in mixed bilayers comprising DPPC and DPPG were studied in a deeper detail upon titration with increasing amounts of DOX (Figure 23). In absence of DOX the thermotropic behavior of DPPC-DPPG at increasing DPPG concentrations in the range of 10 mol% to 30 mol% revealed an ideal mixing, as previously reported [183]. The pre-transition peak at 34.7 °C persisted until 20 mol% DPPG and was abolished at higher concentrations. At various molar ratios of DPPG the binary mixtures exhibited a single thermal transition in the range between the temperatures characteristic for the pure species. With increasing DPPG amount the peak height gradually diminished from 1.515 kW/mol to 0.838 kW/mol, and the main transition peak area also decreased from 37.128 kJ/mol to 29.192 kJ/mol.

Although DOX at a concentration of 20 eq% had little influence on DPPC, the addition of even 10 mol% DPPG strongly increased the effect of DOX on the phase transition. Even at relatively low DOX concentrations – between 0 eq% and 20 eq% according to the molar concentration of DPPG, the pre-transitional peak vanished apparently as a consequence of aglycon penetration between the phospholipid tails.

Figure 23: DCS heat-flow diagrams of the interaction between DOX and anionic DPPC/DPPG phospholipid bilayers (n = 3, second scanning run). Liposomes comprised 10 mol% (A), 20 mol% (B), and 30 mol% (C) of DPPG. In each case the DOX concentration was varied from 0 eq%, 20 eq%, 60 eq%, 100 eq%, 140 eq% and 180 eq% (according to the DPPG amount) at a pH of 7.4. The temperature interval between 25 ºC and 60 ºC is represented. No transition peaks were found out of this temperature span.

Under the same conditions, the main endotherm was broadened and Tm shifted down by 0.67 °C. At DOX concentrations above 100 eq% two overlapping peaks with a Tm difference of about 1 °C could be observed. The higher-melting peak appeared first as a shoulder when 60 eq% DOX were reached and became with increasing drug concentration a separate peak which finally became sharper and centered at about 42 °C. In summary, by raising of DOX concentration three major thermotropic phenomena could be observed:

a) a small reproducible Tm shift towards lower temperatures at low DOX concentrations e.g. the shift from 41.52 °C (0 eq% DOX, 20 mol% DPPG) to 40.85 °C (20 eq%

DOX, 20 mol% DPPG);

b) a gradual decrease of mean transition peak intensity, mostly pronounced in samples with lower concentrations of DPPG e.g. in samples comprising 20 mol% DPPG the

mean peak area (peak 1 at 40.5 °C) decreased from 31.091 kW/mol (0 eq% DOX, 20 mol% DPPG) to zero (140 eq% DOX, 20 mol% DPPG);

c) a gradual formation of a peak shoulder and a new higher-melting peak between 41.42 °C and 41.94 °C at higher DOX concentrations. This effect was the more pronounced, the higher the concentration of DPPG was.

Peak 1 Peak 2

DOX, [eq%]

Tm, [°C] Peak area, [kJ/mol]

Peak height,

[kW/mol] Tm, [°C] Peak area, [kJ/mol]

Peak height, [kJ/mol]

0 41.55 ± 0.11 31.091 ± 0.012 1.394 ± 0.020 --- --- --- 20 40.88 ± 0.05 33.772 ± 0.003 1.072 ± 0.002 --- --- --- 60 41.00 ± 0.08 27.553 ± 0.010 0.811 ± 0.001 shoulder --- --- 100 40.47 ± 0.13 14.366 ± 0.003 0.660 ± 0.005 41.50 ± 0.11 12.008 ± 0.001 0.526 ± 0.005 140 shoulder --- --- 41.42 ± 0.07 7.183 ± 0.011 0.551 ± 0.003

180 --- --- --- 41.94 ± 0.02 26.803 ± 0.007 6.023 ± 0.002

Table 10: Progression of multiple endotherms during titration with DOX of anionic phospholipid bilayers comprising 20 mol% DPPG and 80 mol% DPPC. At low DOX concentration a small Tm shift can be observed, followed by a gradual depletion of peak 1 height. Parallel to that a second higher-melting peak appeared and rose in height at DOX concentrations above 100 eq% (n = 3, m = 3).

Next to the phase transition perturbations in mixed anionic systems, DOX also altered the colloidal and lyotropic condition of liposomes. Small-angle X-ray scattering studies [166]

indicated that DOX in the concentration range of 40 eq% to 80 eq% causes a bilayer reorganization of anionic liposomes into closely-packed multilamellar structures. In this concentration span DOX shows also small but definite fluidizing effects on phospholipid membranes [179] as observable from the peak broadening thermotropic phenomena.

At higher drug concentrations the appearance of high-melting peaks were correlated with the formation of a coarse crystalline phase (Figure 24), exclusion of the aqueous phase, and obliteration of the colloidal system. Surprisingly, although the stoichiometric proportions given by Goormaghtigh et al. (1980) [165] were exceeded, DOX was still bound beyond the concentration mark of 100 eq% (data not shown). Above this limit the concentration-dependent formation of crystalline phase could be related by its optical appearance with the differentiation of a high-melting peak in the thermogram. This finding substantiates the hypothesis that further binding of DOX above the electrostatic charge stoichiometry is possible due to hydrophobic and self-association interactions [166]. Apparently, the excessive

phospholipid complexation enhances phase crystallinity and is detrimental for the liposomes’

colloidal state.

The macroscopic appearance of the crystalline aggregates was as micrometer to millimeter large dark-red particles, prone to sedimentation. A scanning electron microscopy (SEM) image revealed the various form and sharp-edged surface topography of the aggregates (Figure 24). Aggregate crystallinity was confirmed by transmission electron microscopy (TEM, data not shown). The formation of crystalline aggregates and colloidal collapse of the formulations can be explained by escalating formation of stacked aglycon associates [166], interdigitation of phospholipid acyl chains [184] or other unspecific interactions.

Dilution of the anionic phospholipid DPPG with zwitterionic phospholipids such as DPPC lowers the DOX binding ability of liposomes [169], but it also effectively inhibits the formation of crystalline aggregates as seen by the tendency of bilayers with a higher DPPG content to form more intensive high-melting peaks at lower DOX concentrations (Figure 23).

Figure 24: Scanning electron microscopy image of a condensed aggregate of DPPC/DPPG and DOX in formulations, containing 70 mol% DPPC, 30 mol% DPPG, and 140 eq% DOX.

The fluidizing effect of DOX on anionic phospholipid membranes, suggested in the literature [179], has been observed in the above thermograms (Figure 23) in terms of peak broadening and flattening. These peak shape changes can be numerically represented as the ratio between the peak width at half peak height (W0.5) and the peak height (H). For example, in binary mixtures of 10 mol% DPPG and 90 mol% DPPC increasing DOX concentrations cause a continuous increase of the W0.5 / H coefficient i.e. membrane fluidizing until 100 eq%

of DOX are reached (Figure 25, solid line). Beyond this mark the membrane fluidity steeply

declines and crystalline complexes are bound in the final stage. Interestingly, the MB yield follows the same progression as represented by the dashed line plot. This finding together with several further examples later on corroborates the hypothesis that the phospholipid cooperativity is a major factor for the stability of MB shell. This opinion has already been maintained by other authors and witnessed by fluorescence microscopy [14, 79, 93].

Figure 25: Transition peak flattening and broadening (solid line) of peak 1, appearing at about 41°C, in binary mixtures of 10 mol% DPPG and 90 mol% DPPC at DOX concentrations, increasing from 0 eq% to 180 eq%

according to the concentration of DPPG. The dashed line represents the progression of MB yield produced upon mechanical agitation.

The key result of this study was the selection of suitable concentrations for the formulation of DOX-loaded liposomal MB precursors. The choice criteria for the optimum formulation range were: a) the highest molar ratio between DOX and total phospholipid (greatest loading amount of DOX); and b) the highest bilayer fluidity derived from the lowest height of the melting peak, arising below 41 °C (peak 1). The selected interval of 20 mol% DPPG and between 60 eq% and 100 eq% DOX satisfied both requirements, although a better membrane fluidity was existing under 10 mol% DPPG and 100 eq% DOX, yet the drug loading was then lower. Further on in this work, the range of 20 mol% DPPG and 60 eq% to 100 eq% DOX will be used as a milestone for further optimizations and development of DOX-loaded MBs.